JPH02302695A - Updating of parameter of core model and updating of parameter of analytic model - Google Patents

Abstract

PURPOSE: To automatically and periodically renew parameters of an analytical core model by comparing computed values relevant to parameters of the core model with measured values relative to parameters computed from systematic values and adjusting these values. CONSTITUTION: A model regulation system 10 receives various kinds of plant instrumentation data from a nuclear reactor controlling system 12, a control rod controlling system 14, a thermocouple system 16, a nuclear reactor protection system 20, and a detector system 24 movable in the reactor and normally these systems provide software modules of a plant computer. All of the measurement instruments repeatedly receive equilibrium outputs and other model data from model files, the elements of the computed output distribution are compared with corresponding elements of outputs in parallel axial direction obtained from a conventional equilibrium flux map and adjusted, and the critical output shape computed from the model files is conformed to the measured output. After that, the regulation system 10 receives plant instrumentation data through a processor and renews model files and at the same time approximately continuously and automatically regulates the analytical core model in the on-line manner.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is a system for automatically and periodically updating and adjusting a core analytical model of an online pressurized wooden reactor consisting of a data file of parameters representing the core, At all times, it ensures that the model (data file) closely matches the current characteristics of the core being modeled and monitored, and provides plant personnel with a real-time, predictive representation of the operating characteristics of the actual core. The attached core parameter prediction device can be displayed for
It allows users to reliably initialize at any time, even when only minimal core monitoring analytical capabilities are available.

In the present invention, the analytical core model has three essential elements. The first element is an extensive collection of facts, including a description of the core's physical content and nuclear cross-sections representing the relative rates at which various nuclear reactions occur within the core. The second factor is the combination of the spatial distribution of time-varying concentrations of current transient nuclear isotopes, which has a substantial effect on the local neutron balance throughout the core. Related representative isotopes are xenon-135 and samarium-149 precursors, iodine-
135 and promethium-149, and long-term combustion. The third essential element of this model is a small set of coefficients that work together with the algorithms applied to the neutron calculation procedure to account for the complex nuclear phenomena known to occur in reactor cores. It is reproduced with considerable accuracy using an extremely simple approximation method.

The concept of updating the analytical core model refers to the temporal tracking or follow-up monitoring of changes in the local concentration of some nuclides that are particularly important in satisfying the second element of the analytical core model. The concept of analytical core model tuning constitutes the third element of an analytical core model, so that simple approximations used to reproduce the effects of complex nuclear processes in the core have the most effective reproducibility. It refers to changing one or more coefficients. There are two unique problems with the above elements.

The first problem is to provide reliable and concise instructions to reactor operators about both the actual current core status and the current core status trends.
To enable reactor operators to effectively and efficiently operate control functions using xenon distribution displays such as those described in U.S. Pat. No. 4,642.213;
Regarding the need to obtain sufficient online information related to the extraction power, current distribution of iodine-135 and xenon-135. In this regard, reactor operators may well include dedicated automated control systems that nominally perform human control functions. There are three known means to satisfactorily solve this problem. The first is to utilize the responses of a large number of fixed in-core detector chains to synthesize a complete three-dimensional core power distribution from which the required indications can be easily derived; is to use an online three-dimensional analytical core model augmented with relatively few plant instrumentation signals to form a relatively useful three-dimensional core power distribution.

Furthermore, a third means describes the detailed one-dimensional core average axial power distribution and US Pat. Uses only signals from conventional plant instrumentation filtered through empirically derived correlations to produce the required derivative indications as described in '774,050. That is what it is. The use of multiple fixed in-core detector chains necessarily imposes relatively high equipment installation and installation costs on the plant owner. The use of three-dimensional analytical models significantly increases computer resources. Using purely empirical correlations requires constant careful calibration of the plant instrumentation and the correlation itself.

The second problem is that the analytical core Regarding the need to ensure that the model is fairly well matched to the operating characteristics of the corresponding core. When the plant is operating, certain operating parameters of the pressurized water reactor (power level, control bank level 1 (Con
Trolbank position, etc.) and to periodically update the analytical core model by actually creating images of the core response when the core is actually reacting. In the past, various attempts have been made using conventional core models. In these cases, there was no attempt to adjust any of the coefficients of the third element of the core model on-line to force the core model to match the actual nuclear properties of the core. In all such cases, where the reactor involved is unstable or slightly unstable with respect to spatial xenon oscillations, the calculated core average axial power fraction and the actual measured core average axial power distribution Deviations or deviations (as inferred by comparing the calculated axial offset with the measured offset) between the , making subsequent predictions of core response questionable. This second problem requires slightly more computer resources, but is at least slightly coupled to the plant instrumentation and allows for periodic adjustment of the actual neutron properties of the model to measurable values of the core properties. This can be solved by using a full three-dimensional nodal core model that can reasonably match the core characteristics calculated by Notwithstanding the foregoing, it is clear that what is needed is a simple online-dimensional analytical core model. That is, this model requires far fewer computer resources than a fully three-dimensional analytical core model. The actual neutron characteristics of the core model can be adjusted as deemed necessary, ensuring that the calculated axial power distribution continues to closely track the measured axial power distribution.

To continuously update the distribution of axial power, iodine, xenon, promethium, samarium and long-term combustion of the core model, and to simultaneously adjust the nuclear properties of the model to match the nuclear properties of the core. , a close match of the core model to the actual reactor obtained by using monitored reactor instrumentation responses is a relatively inexpensive and easily implemented method for solving the two problems mentioned above. I will provide a.

U.S. Pat. No. 4,711,753 discloses results obtained from an equilibrium complete core flux map for calibrating or adjusting elements of an analytical model (data file) used by a core response predictor. A proposal to use the has been disclosed. More specifically, the lateral buckling value (trans
The axial distribution B2xy(z) of reverse buckliB values) is adjusted so that the calculated axial power distribution of the core model closely approximates the core average axial power distribution obtained from the bundle map. Since information about the transient iodine, xenon, promethium or samarium distribution cannot be obtained from a single bundle map, the condition is imposed that the east map must be obtained only under stable equilibrium core conditions. In such means, the axial distribution of the lateral buckling value is expressed by the following equation.

This formula is expanded as follows.

B2. , (z) = to −F, (2) to ten, F2 (2) to thousand, F, <z>÷＾, F, (z) + ＾5Fs(z) ten・
...(2) The calibration process or adjustment process is an expansion coefficient formula. , which is a set of integral parameters characterizing the calculated axial power distribution.
ara-meters) in good agreement with the corresponding parameters characterizing the core average axial power distribution obtained from the bundle map. The specific parameters are the well-known axial offset parameter (80) and one third of the core height.
The complex relationship of the calculated axial power distribution to the core model (of which lateral buckling is only one component), including higher order terms representing quarter and fifth integrals. Therefore, the value of the expansion coefficient is determined by several nested levels of search (nested levels of 5earc).
h) A guided trial and error search process (guided trial and error search process)
trial and error search pr
The specific function F,(z) in the expansion used, which has to be found by the The whole procedure can be carried out because it only has the property of decoupling. In this way, the coefficient ^1 has an effect on the reactivity balance, but it has almost no effect on the output distribution, and the coefficient ^2 limits the axial offset of the output distribution, but it has no effect on the axial pinch (AP). The effectiveness of decoupling, which does not substantially affect ing. However, U.S. Patent No. 4,
The calibration procedures described in '711,753 are applicable only if new flux map results obtained at equilibrium core conditions are available. This typically occurs once a month in conventional operating pressurized water reactors.

Therefore, the system required is capable of handling at least major lateral buckling until the next monthly calibration is available to compensate for minor deficiencies in the analytical core model and/or the neutron algorithms used. It must be possible to readjust the values of the coefficients, especially the coefficients ^2,^ of equation 2) on-line and on a nearly continuous basis.

11.1 It is an object of the present invention to provide a method for continuously updating parameters of the current distribution and associated nuclear reactor core analytical model.

It is also an object of the present invention to provide a robust representation of the actual characteristics of the monitored core, so that other initialization parameters of the analytical core model can be updated accurately and effectively during reactor operation. The objective is to provide a method for continuously adjusting the contents of an analytical core model on-line in order to guarantee the accuracy of the analysis.

Yet another object of the present invention is to provide update and adjustment functionality without the need for complex core power distribution synthesis functionality.

It is another object of the present invention to provide a system for updating and adjusting the model without requiring user intervention so that the model is accurate when needed.

The above objectives can be achieved by a system that periodically uses available measurements of conventional core instrumentation to update and adjust an analytical core model. An initial baseline calibration of the core model is performed using the results from the equilibrium flux map and the simultaneously measured values of the reactor cooling system boron concentration. After this, the analytical model is updated and adjusted online to reproduce the actual core behavior during core operation. To accomplish this, changes are made to the model to track measured changes in core power level, control bank position, core inlet temperature, etc., and the model is updated in progressive short steps with each update. Update calculated values of depleted axial power distribution and axial distribution of iodine, xenon, promethium, samarium and long-term combustion. At the same time, the parameter values of axial offset and axial pinch are taken from the calculated axial power distribution and the actual values obtained directly from the conventional core instrumentation response, without the need to perform intermediate core averaged axial power distribution synthesis. are compared with the estimated values of the core average axial offset and axial pinch parameters to determine whether adjustments to the analytical core model coefficients are necessary. If the parameters of the calculated axial offset or the calculated axial pinch differ from the measured values of the axial offset or axial pinch by more than a predetermined tolerance limit, the analytical lateral buckling deployment formula An adjustment step regarding the second and third coefficients of is performed.

These and other objects and advantages will become apparent from the structure and operation described in detail below, and reference is made to the accompanying drawings in the following description.

In addition, in the figures, the same reference numerals indicate the same parts.

The Japanese invention, when used on a nearly continuous basis and online, requires frequent periodic updating of part of the contents of a -dimensional (axial) analytical model and, if necessary, updating of other parts of that model. im of the content of the reactor, allowing the analytical core model to perform two functions: The second function is to supply the graphical system with the necessary data that can form a graphical display related to current core conditions and trends; To provide a reliable basis for initiating a series of analytical predictions of the intended core response for a projected plant plan, when required by the system.

Furthermore, the present invention eliminates the aforementioned tendency for the analytically updated axial power distribution produced by the core response predictor to gradually deviate from the true core average axial power distribution. Forcing the analytically calculated axial power distribution to be very close to the true core average axial power distribution also allows the calculated axial power distributions for iodine, xenon (promethium and samarium) and long-term combustion Guarantees close approximation to the current core average distribution.

In particular, the present invention provides "measured" values of in-core axial offsets synthesized from conventional plant instrumentation response signals on a continuous, automatic and periodic basis.
The axial offset of the axial power distribution determined and calculated as a basis for adjusting the coefficient ^2 in the analytical formula of the axial distribution of the lateral buckling value of the analytical core model;
Match the measured in-furnace axial offset. The present invention also synthesizes in-core axial pinch measurements from plant instrumentation response signals, and the measured axial pinch values are identical to the axial distribution of lateral buckling values in the analytical core model. used to adjust the coefficient ^3 in the display formula,
The calculated axial pinch of the axial power distribution is matched with the measured in-furnace axial pinch value.

In addition to measuring the cold leg temperature, control bank position, and core power level currently supplied to the existing core prediction system, the present invention also measures the upper detector of at least one conventional ex-core power range neutron detection channel. and the signal from the bottom detector at a minimum. A signal from at least one particular core inlet thermocouple is also added to this input signal combination. Thermocouple selection is based on U.S. Patent No. 4,774
, 050, and which also includes one or more so-called "a+ulti-section
)” If an in-reactor neutron detector is used in a particular instrumentation, or one or more fixed in-core detector chains (with at least three detectors in one chain) If used, the signals from the individual ex-core detector sections or fixed in-core detectors replace the combination of the two section in-core detector signals and the core exit thermocouple signal.The two detector signals is used to synthesize the measured axial offset value (^0) in conjunction with the control bank position signal and other variables fed into the model by directly applying the following equation.

Here, Bl to B3 are a set of transient bundle maps (trans
ientNux map) to the least squares fit calibration method (1east-5quares fit calibr
where DRL, DR, are the signals from the upper and lower parts of the in-reactor detector, and Q is the core thermal power normally given to the prediction module. level and bp is the control bank position. In the least-squares fit calibration method described above, a series of bundle maps are taken for several days every three months when the core is under transient axial vibration conditions, and the bundle maps are taken at various axial offsets. A least squares fit is taken to obtain the coefficients. The axial pinch is synthesized by adding a term representing the local coolant enthalpy increase in the peripheral region of the core as detected by in-core detectors and using the following correlation:

where CI-C4 is the axial pinch coefficient obtained by a least squares fit calibration method against a series of transient flux maps as described above, and Δh is the
The local enthalpy increase obtained from the core exit thermocouple and from the cold leg temperature signal as described in the '000 patent.

If one or more "multi-section" in-core detectors or one or more fixed in-core detector chains are available, the measured axial offset value and the corresponding measured axial pinch value are determined by the following equation: Synthesized using

801144g””FAQ (ILOP+l1lj#@r
@1ddl*+I lower m1dd111+
1110uLO@+hD) (5) 8P 11611
M”FAP(ItOj+l1ij#@r ald414
+11(III@r @1441@+Ib@tLOm+
hD) (6) where ■ represents the current generated by the individual parts of a "multi-part" in-core detector, for example, and the function is a simple numerical correlation function between the value of the actual offset or pinch detection i defined by said value, which can be determined by a nuclear engineer.

The present invention provides a method for controlling the time of day, power level, control bank position, cold leg temperature, RC3 pressure (optional), axial offset, and axial pinch values on a prespecified periodic basis (e.g., every 5 minutes). Determine. time, output level,
Control bank position, cold leg temperature, and pressure (if provided) are used to update a model data file containing current core model parameter content.

The core axial power distribution calculation values for axial offset and axial pinch can be obtained from the model. The calculated values of axial offset and axial pinch are compared with the measured values of axial offset and axial pinch. If the respective values agree within pre-specified tolerance limits, i.e. ,
In addition to a more accurate follow-up monitoring function, if the current deviation or time-integrated deviation is below a predetermined threshold, the updated model core contents are memorized and the update process is repeated at the next predetermined update time. will be interrupted until If the value of the calculated axial offset or the calculated axial pinch does not match the corresponding measured value within the specified tolerance, then US Pat.
．． Similar to No. 711,753, the coefficient ^2 of the analytical expression for the axial distribution of lateral buckling values is used to provide acceptable agreement between the calculated and measured values of axial offset and axial pinch. , ^, are adjusted. If a satisfactory match is obtained, the resulting analytical core contents are stored and the update process is suspended until the next predetermined update time.

As shown in FIG. 1, a model adjustment system or model adjustment system 10 receives core thermal power from a reactor control system 12 and obtains control bank position information from a control rod control system 14. A thermocouple system 16 connected to a thermocouple 18 placed at the outlet of the fuel assembly in the reactor core measures the cold leg temperature obtained from the reactor control system 12 as well as the increase in enthalpy.
In addition to providing measurements, a neutron detector signal is obtained by a reactor protection system 20 connected to an in-core detector 22 or by an in-core fixed detector system 24 connected to an in-core detector chain 26. . System or system 10.12.14.16.
20.24 is typically provided as a software module on a plant computer.

The relationship of the model adjustment system 10 to other software modules is as shown in FIG.
r) iteratively receives the equilibrium output, axial iodine, xenon, promethium, samarium and long-term burn up distributions along with the input boron concentration from the model file 42 and uses a parallel xenon distribution. The conventional one-dimensional diffusion theoretical formula for the axial output shape is executed, and the elements of the calculated output distribution are ^0, AP, 80, A.
The coefficients of equation (2) are compared with the corresponding elements of the average axial power distribution obtained from a conventional balanced flux map,
is adjusted to closely match the calculated critical output shape of model 42 to the measured output shape as shown in U.S. Pat. No. 4,711,753.

Thereafter, a conventional front-end data processor 44 receives the aforementioned plant instrumentation data and provides it to the regulator module 10. Simultaneously with the update of the model 42 by the model update system 46, the analytical core model 42 is automatically and continuously adjusted on-line. With a substantially continuously adjusted analytical core model as a starting point for predictions in response to a user's specific plan, a conventional core response predictor 48 uses the axial iodine, xenon, promethium , samarium, long-term burn-up, and lateral buckling distributions can be used to predict the reactor's response to expected changes entered by the user.

Additionally, this model may be used in a conventional graphical display generator 52, such as that described in U.S. Pat. No. 4,642,213, to form a display for a user 50 on a graphical monitor 54. I can do it.

An example of the process steps required in the present invention to automatically and continuously update the analytical core model starting parameters without user intervention is shown in Figure 3 every six months. At the beginning of the update cycle of (step 60),
A bundle map is conventionally obtained, the boron concentration in the reactor cooling system is measured, a calibration is performed as described in U.S. Pat. No. 4,711,753, and a model or data file is created. The coefficients ^1 to ^5 are calibrated (step 62), and, less preferably, the coefficients ^1 to ^5 are periodically adjusted from the reactor design formula.
can be obtained. Once the model is calibrated, or a determination is made that monthly calibration is not required, the current state of the core is read progressively as core operations progress (step 64). This reading includes obtaining time of day, power level, control rod position, inlet temperature, pressure, detector readings, and core exit thermocouple readings from core instrumentation sensors. In this step, the core model 42 is then read from its memory location and the core model is
Updated using conventional impairment calculations using power level, control rod position, population temperature and pressure (step 66)
.

The current analytical axial offset and axial pinch values are then found in the full range calibrator 40 or core response predictor 48 by performing a conventional criticality search using a one-dimensional diffusion theory algorithm. , the current actual axial offset and pinch values calculated using conventional neutron equations such as the -dimensional nodal (nodat) algorithm or the -dimensional neutron transfer algorithm, as described in U.S. Pat.
, 774゜050, when the enthalpy increase Δh is calculated, equation (3) (
4) using the detector readings, thermocouple readings, control rod position, inlet temperature and power level (step 70). The actual value determined using equations (3) and (4) is compared with the analytical value (step 72), the absolute difference between the calculated analytical axial offset and the measured actual axial If the values differ by less than a predetermined value n, e.g. by less than 0.5%, and the absolute value of the difference between the calculated analytical axial pinch value and the measured actual axial pinch value If the difference is less than a predetermined value m, for example also less than 0.5%, then these differences are acceptable and the model II! A more accurate method, which does not require any adjustment, is to use the above absolute value as
A time integration should be performed before each comparison is made with a predetermined constant Ω, m, which allows a more accurate determination of the total or integrated drift in the core from the previously estimated model starting values. can.

That is, the increase or decrease in the integral value of the axial offset and axial pinch deviation accumulated since the last re-adjustment of the parameters ^2,^, should be checked against predetermined limits. be.

If the comparison is made and the limit is found not to be exceeded, the system stores the model or data file (step 74) and waits for a predetermined period of time (step 76). That is, it remains inactive until the time of the pond's periodic update cycle. As shown in dotted box 75, the system may also display the adjusted and updated model by inputting the model to display generator 52. If the deviation or shift is significant, i.e. unacceptable, the value of the factor ^2,8 in the buckling equation is adjusted (step 78).
).

At these adjusted coefficient values, the neutron equation is also used to determine newly calculated axial offset and axial pinch analytical values (step 80);
The cycle of comparison 72, adjustment 78 and calculation 80 is the coefficient ^
2, ^3 is the non-zero reverse sign deviation value (non-zero o
pposite sign deviati'on v
The standard method for generating alues is carried out periodically until compensated for the drift using the so-called overcompensation in control theory. This type of compensation method requires that a deviation be canceled by a backward deviation that is slightly smaller than its magnitude, e.g., if the deviation is positive in the positive direction.
If calculated as 5%, the compensation criterion requires that the compensated result be shifted in the negative direction by a value of, for example, 0.25%. The reason for overcompensation is that the error that accumulates in the axial distribution of iodine and xenon is actually the time integral of the axial power distribution error, so overcompensation effectively forces the system to and eliminates xenon errors.

The time constants of promethium and samarium are relatively large compared to iodine and xenon, so their errors are relatively insensitive to variations in the output distribution error, and when compensation is done, the model is memorized. (step 74), wait for an interruption to occur (step 76) and start a new cycle;
Alternatively, the display is performed before the interruption step 76 (step 75).

In addition to being able to ensure that the prediction of core response to a scheduled plant operating plan starts from real, current initial conditions, several other advantages derive from the effectiveness and use of the present invention. Since the present invention is carried out continuously and online in a specific facility, changes in the values of the monitoring operating parameters supplied to the system and the coefficients ^2,^ of the expansion formula are recorded in the process. be able to. Such a recording step can be set after step 74. This record is then analyzed offline, either manually or automatically, to identify changes in operating conditions and parameter values.
We can find a correlation between the adjustment of , ＾, °, f! Any mutual rWJgA recognized would directly demonstrate shortcomings of conventional computational neutron algorithms; such shortcomings would later be corrected by analysts, so that
Analytical algorithms gradually come to match reality.

The calculated critical reactor cooling system boron concentration is obtained as a by-product of the criticality search. This is because the boron concentration is recalculated during updates and adjustments. As a result, when a reliable value for the boron concentration in the reactor cooling system is obtained, the coefficient will also be adjusted. Because the core model is routinely adjusted to match concentrations, and whenever suitable measurements are obtained, the model is revised to match the calculated boron concentration to the measured values. A reliable estimate of the boron concentration of can be displayed for use by personnel. Furthermore,
Systematic errors in boron concentration calculations from measurements can indicate shortcomings in analytical core models.

Because the analytical core model is checked against the plant just before any trip that may occur, the invention also allows the model to be updated when the core is in a subcritical state (although it cannot be confirmed). , the present invention can be modified to process estimated critical states and outage margin estimates without much intervention by the user, other than outputting the results.

The many features and advantages of the invention will be apparent from the foregoing detailed description, and the appended claims will reveal all such features and advantages as come within the true spirit and scope of the invention. Furthermore,
It is not desirable to limit the invention to the precise construction and operation described above, as many modifications and variations will readily occur to those skilled in the art, and all suitable modifications and equivalents may be made without departing from the scope of the invention. It's good to judge.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram showing a nuclear reactor having a model adjustment system or system according to the present invention; FIG. 2 is a schematic diagram showing the relationship between the model adjustment system of the present invention and other systems related to determination and prediction of core state; FIG. 3 is a schematic diagram showing the operation of the system of the present invention. In the figure, 10...Model adjustment system 12...Reactor control system 1
4... Control rod control system 16... Thermocouple system 18...
・Thermocouple 20...Reactor protection system 22...
In-furnace detector 24... in-furnace fixed detector system FIG,
1 Plant data FIG, 2

Claims (1)

[Claims] 1. A method for updating parameters of a core model for a predetermined system, comprising: (a) calculating calculated values related to parameters of the core model; and (b) measurement. (c) determining a deviation between the calculated value and the measured value; (d) the deviation exceeds a predetermined value; (e) updating the parameters of the core model when the parameters are compensated. 2. A method for updating parameters of an analytical model used for predicting distribution in the core of a nuclear reactor, comprising: (a) periodically calibrating the analytical model; and (b)
(c) updating the analytical model using an impairment formula; (d) axial offset and axis; (e) calculating actual axial offset and axial pinch measurements from the enthalpy rise, reading, output level, position, and axial offset/pinch expansion equation coefficients; (f) integrating the difference between the measured value and the analyzed value; (g) comparing the absolute value of the difference with a predetermined compensation indication limit; and (h) the absolute (i) adjusting second and third buckling coefficients until said measured value is compensated if one of the values is greater than a corresponding limit value; and (i) if said calculated value is compensated; A method for updating parameters of an analytical model, comprising: updating the parameters; and (j) displaying an analytical model including the updated parameters.